What Vaire does

Vaire is a UK-headquartered semiconductor startup developing adiabatic reversible computing hardware designed to recover and recycle energy rather than dissipate it as heat. The company's core technology uses reversible logic gates that allow computations to run in both directions, enabling input information to power subsequent calculations instead of being emitted as waste heat. Vaire's first silicon is the Ice River test chip, fabricated in 22nm CMOS, which in September 2025 demonstrated net energy recovery with factors of 1.77 for a capacitor array and 1.41 for a shift register, alongside approximately 50% energy savings in initial AI workload tests.

The company targets AI infrastructure and data center operators as its primary customer segment, positioning its architecture against the unsustainable energy demands of classical CMOS chips and the diminishing returns of Moore's Law scaling. Co-founded in 2021 by CEO Rodolfo Rosini and CTO Dr Hannah Earley (PhD, Applied Maths & Theoretical Physics, University of Cambridge), Vaire has assembled a leadership team drawing on talent from Arm (founding engineer Andrew Sloss as VP Technology), Intel Foundry and Intel Labs (Steve Pawlowski as Fellow, former lead of Intel's Reversible Computing team), Sandia/MIT/Stanford (Senior Scientist Mike Frank), and senior operating experience from Google, Mapbox, and Grail (Interim COO Paige Zeigler). The company operates from London with additional offices in Bellevue and Sunnyvale, and has been backed by 7percent Ventures, Seedcamp, Lifeline Ventures, and Intel Ignite.

Vaire is pre-revenue and remains in the prototype stage with no commercially available products, no published pricing, and no announced distribution channels. Its go-to-market motion is still forming: the company has so far relied on earned media in technology and science publications (Science News, American Institute of Physics, Jon Peddie Research), whitepapers, podcast appearances, and organic social channels to build awareness. Revenue mechanics are undefined, and the path to commercialization will require transitioning from the Ice River research demonstration to a production-grade resonator chip suitable for AI inference deployment.

Name

Vaire

Legal name

Vaire Computing Inc.

Website

https://vaire.co

Company type

Private

Founded year

2021

Operating status

Operating

Headcount range

1–10 employees

Short description

Vaire is a UK-based semiconductor startup developing adiabatic reversible computing chips that recycle energy instead of dissipating it as heat. The company targets AI infrastructure and data center workloads, validating proof-of-concept with its 22nm Ice River test chip in September 2025.

Ownership category

akta.pro rank

Headquarters

HQ city

London

HQ country

United Kingdom

HQ region

Europe

Offices3 records

Markets served

GTM type

B2B

Offering type

Hardware or Manufacturing

Cost components

Technology or R&D, Personnel, Operations, Infrastructure, Marketing or Sales

Marketing channels5 records

Core offering

Vaire is developing near-zero energy computing hardware based on adiabatic reversible computing. Its core offering is the Ice River test chip, built in 22nm CMOS, which validates reversible logic that recovers and recycles energy rather than dissipating it as heat, and is intended to commercialize resonator-based chips targeting AI inference workloads. The company is currently pre-revenue and in the prototype/validation stage.

Product overview

Vaire is developing near-zero energy computing hardware based on reversible computing technology. The company's first product is the Ice River test chip (codenamed Ice River), built in 22nm CMOS, which demonstrates adiabatic reversible computing that recovers and recycles energy rather than dissipating it as heat. The technology targets AI inference workloads, where the company has demonstrated approximately 50% energy savings compared to traditional chips. Vaire's reversible computing approach enables computing in both directions, recycling input information to power new calculations rather than emitting it as heat, offering a path to decouple computing power growth from energy consumption increases.

Differentiator

Problem solved

Functional benefit

Products and services

• Ice River Test Chip A first-of-its-kind adiabatic reversible computing test chip built in 22nm CMOS, demonstrating net energy recovery with factors of 1.77 (capacitor array) and 1.41 (shift register) versus traditional square-wave driven circuits. It validates the reversible computing proof of concept for AI inference workloads and underpins Vaire's planned commercialization of resonator-based near-zero energy chips.

Quantifiable outcome

• 50% energy savings in initial tests of reversible computing chip for AI
• +2 more outcomes

Segments1 record

Ideal customer profiles1 record

Technology focussed Yes

API detail

Has API

No

API docs

API detail

Core technology

AI maturity

App detail

AI capability1 record

Feature3 records

Recent moves6 records

Expansion highlights5 records

Direct peers

• Groq: Groq builds a custom LPU inference accelerator targeting the same AI inference and low-latency workload segment as Vaire. Both companies position novel non-NVIDIA silicon for high-throughput, low-latency AI serving in data centers.
• Lightmatter: Lightmatter develops photonic compute hardware for AI workloads, taking a fundamentally novel compute substrate (photons instead of electrons) to attack the same energy-efficiency bottleneck Vaire targets with reversible/adiabatic CMOS.
• Cerebras Systems: Cerebras builds wafer-scale AI compute systems targeting inference and training workloads. Like Vaire, it pursues a non-GPU, novel-architecture path to differentiate on throughput and efficiency for AI customers.
• SambaNova Systems: SambaNova develops the RDU dataflow architecture for enterprise AI training and inference. It competes in the same custom-AI-silicon-for-data-centers category as Vaire with a similarly novel compute architecture thesis.
• Tenstorrent: Tenstorrent, led by Jim Keller, designs RISC-V-based AI accelerators (Grayskull/Wormhole/Blackhole) targeting inference and training. Like Vaire, it sells novel AI compute hardware to displace NVIDIA in data center deployments.
• Etched: Etched builds the Sohu ASIC, a transformer-specific AI inference chip. It is a direct comparable to Vaire as a venture-backed startup designing a custom accelerator from first principles for AI workloads.
• d-Matrix: d-Matrix builds in-memory compute (DIMC) AI inference accelerators targeting data center LLMs. It is a comparable AI-inference-focused chip startup pursuing a fundamentally different compute substrate to attack energy efficiency.
• Positron AI: Positron develops custom AI inference silicon (Atlas) for transformer workloads. Like Vaire, it is a venture-stage startup offering an alternative to NVIDIA GPUs for data center inference customers.

Broad incumbents

• NVIDIA: NVIDIA is the dominant incumbent in AI compute (H100/B200 GPUs and CUDA software). Vaire's reversible compute approach is explicitly positioned as an alternative path to delivering AI performance, making NVIDIA the primary competitive benchmark.
• Intel: Intel Labs has historically hosted the most prominent industrial reversible computing research program (formerly led by Vaire Fellow Steve Pawlowski). Intel Foundry also manufactures advanced-node silicon, positioning Intel as both a potential competitor and a future foundry partner.

Market position

Strengths5 records

Weaknesses5 records

Competitive moat3 records

Key risks6 records

Key highlights5 records

Customer concentration

• Xx.com/VaireHQ
• LinkedInlinkedin.com/company/vaire-computing

Revenue estimate

Valuation estimate

Number of profiles

Profiles6 records

Funding overview

Funding rounds4 records

Investors5 records

Funding detail is available on the Subscription and Enterprise plan.Contact sales →

M&A

Investments

M&A and investment is available on the Subscription and Enterprise plan.Contact sales →

Frequently asked questions about Vaire